def calc_by_states(df):
"calc parameters by state"
plotdir = make_plotdir(plotdir='cms_state_service_plots/')
agg_fns = ['count','median']
p_group = calc_par_group(df, agg_fns, ['provider_type','nppes_provider_state'], ['pay_per_person','pay_per_service'], print_out=False)
# print('index level provider_types\n', p_group.index.levels[0])
bmap = get_basemap() # read file once for all maps
# minmax = (1.0, 2.7)
# im = p_group.ix['Internal Medicine']['pay_per_service']['median']
# print('%s\n' % provider, im)
# make_state_map(bmap, im, plotdir, 'cost_per_service_internal_medicine', 'Internal Medicine, Median Cost Per Service')
# im = p_group.ix['General Surgery']['pay_per_service']['median']
# print('%s\n' % provider, im)
# make_state_map(bmap, im, plotdir, 'cost_per_service_general_surgery', 'General Surgery, Median Cost Per Service')
# im = p_group.ix['Physical Therapist']['pay_per_service']['median']
# print('%s\n' % provider, im)
# make_state_map(bmap, im, plotdir, 'cost_per_service_physical_therapist', 'Physical Therapist, Median Cost Per Service')
patr = re_compile('[ (/)]+')
# for provider in p_group.index.levels[0]:
# im = p_group.ix[provider]['pay_per_service']['median']
# make_state_map(bmap, im, plotdir, 'cost_per_service_%s' % '_'.join(patr.split(provider.lower())), '%s, Median Cost Per Service' % provider)
plotdir = make_plotdir(plotdir='cms_state_person_plots/')
for provider in p_group.index.levels[0]:
im = p_group.ix[provider]['pay_per_person']['median']
make_state_map(bmap, im, plotdir, 'cost_per_person_%s' % '_'.join(patr.split(provider.lower())), '%s, Median Cost Per Person' % provider)
def gender_par_groups(df):
"calculate series of grouped gender parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_pop_gender_plots/')
agg_fns = ['count','sum']
pars = ['beneficiary_female_count','beneficiary_male_count']
p_group = calc_par_group(df, agg_fns, ['provider_type'], pars)
make_group_bar_plots(p_group, 'provider_type', pars, ['female','male'], 'sum', 'Patient Gender', plotdir)
def pop_calc_par_groups(df):
"calculate series of grouped population parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_pop_plots/')
agg_fns = ['count','sum','median']
p_group = calc_par_group(df, agg_fns, ['provider_type'], ['total_unique_benes','total_services','beneficiary_average_age','total_medicare_payment_amt'])
print('\ntop total_services count') # count of number of providers, not patients
p_sort = filter_group_by_var(p_group['total_services'], agg_fns, stat='count')
make_bar_plot(get_col(p_sort,'count',log=True), plotdir, 'total_services_count', 'Log10 Count Total Services', xlim=(0.1,5))
print('\ntop total_services sum')
p_sort = filter_group_by_var(p_group['total_services'], agg_fns, stat='sum')
make_bar_plot(get_col(p_sort,'sum',log=True), plotdir, 'total_services_sum', 'Log10 Sum Total Services', xlim=(1,9))
print('\ntop total_unique_benes sum')
p_sort = filter_group_by_var(p_group['total_unique_benes'], agg_fns, stat='sum')
make_bar_plot(get_col(p_sort,'sum',log=True), plotdir, 'total_unique_benes_sum', 'Log10 Sum Total Beneficiaries', xlim=(1,8))
print('\ntop beneficiary_average_age median')
p_sort = filter_group_by_var(p_group['beneficiary_average_age'], agg_fns, stat='median')
make_bar_plot(get_col(p_sort,'median'), plotdir, 'beneficiary_average_age_median', 'Median Beneficiary Age Per Provider', xlim=(50,100))
print('\ntop total_medicare_payment_amt median')
p_sort = filter_group_by_var(p_group['total_medicare_payment_amt'], agg_fns, stat='median')
make_bar_plot(get_col(p_sort,'median',log=True), plotdir, 'median_medicare_payment_amt', 'Log10 Median Medicare Payment Amount Per Provider', xlim=(1,6))
print('\ntop total_medicare_payment_amt')
p_sort = filter_group_by_var(p_group['total_medicare_payment_amt'], agg_fns, stat='sum')
make_bar_plot(get_col(p_sort,'sum',log=True), plotdir, 'total_medicare_payment_amt', 'Log10 Total Medicare Payment Amount', xlim=(1,10))
def gender_par_groups(df):
"calculate series of grouped gender parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_pop_gender_plots/')
agg_fns = ['count', 'sum']
pars = ['beneficiary_female_count', 'beneficiary_male_count']
p_group = calc_par_group(df, agg_fns, ['provider_type'], pars)
make_group_bar_plots(p_group, 'provider_type', pars, ['female', 'male'],
'sum', 'Patient Gender', plotdir)
def pay_calc_par_groups(df):
"calculate series of grouped pay parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_cost_plots/')
agg_fns = ['count','median','mean','std']
p_group = calc_par_group(df, agg_fns, ['provider_type'], ['pay_per_person','pay_per_service'])
print('\ntop pay_per_service')
p_sort = filter_group_by_var(p_group['pay_per_service'], agg_fns, stat='median')
make_bar_plot(get_col(p_sort,'median'), plotdir, 'pay_per_service', 'Median Log10 Pay Per Service')
print('\ntop pay_per_person')
p_sort = filter_group_by_var(p_group['pay_per_person'], agg_fns, stat='median')
make_bar_plot(get_col(p_sort,'median'), plotdir, 'pay_per_person', 'Median Log10 Pay Per Person')
# p_sort = filter_group_by_var(p_group['pay_per_person'], agg_fns, stat='count')
make_bar_plot(get_col(p_sort,'count'), plotdir, 'pay_per_person_count', 'Count Per Person')
make_scatter_plot(get_col(p_sort,'count'), get_col(p_sort,'median'), plotdir, 'pay_per_person_by_count', 'Count', 'Log10 Cost Per Person', xlim=(-1000,100000))
plotdir = make_plotdir(plotdir='cms_gender_plots/')
g_group = calc_par_group(df, agg_fns, ['provider_type','nppes_provider_gender'], ['pay_per_person'])
g_group = process_by_var(plotdir, g_group, col='pay_per_person', var='nppes_provider_gender')
def average_age_par_group(df):
"average age calc per provider type"
df['total_age'] = df['beneficiary_average_age'] * df['total_unique_benes']
# print(df['total_age'])
plotdir = make_plotdir(plotdir='cms_pop_plots/')
agg_fns = ['sum']
p_group = calc_par_group(df, agg_fns, ['provider_type'], ['total_unique_benes','total_age'])
p_group['avg_age'] = p_group['total_age']['sum'] / p_group['total_unique_benes']['sum']
p_sort = p_group.sort_values(by='avg_age', ascending=False)
print_all_rows(p_sort, ['avg_age'])
make_bar_plot(get_col(p_sort,'avg_age'), plotdir, 'beneficiary_average_age', 'Beneficiary Average Age', xlim=(50,100))
def pay_calc_par_groups(df):
"calculate series of grouped pay parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_cost_plots/')
agg_fns = ['count', 'median', 'mean', 'std']
p_group = calc_par_group(df, agg_fns, ['provider_type'],
['pay_per_person', 'pay_per_service'])
print('\ntop pay_per_service')
p_sort = filter_group_by_var(p_group['pay_per_service'],
agg_fns,
stat='median')
make_bar_plot(get_col(p_sort, 'median'), plotdir, 'pay_per_service',
'Median Log10 Pay Per Service')
print('\ntop pay_per_person')
p_sort = filter_group_by_var(p_group['pay_per_person'],
agg_fns,
stat='median')
make_bar_plot(get_col(p_sort, 'median'), plotdir, 'pay_per_person',
'Median Log10 Pay Per Person')
# p_sort = filter_group_by_var(p_group['pay_per_person'], agg_fns, stat='count')
make_bar_plot(get_col(p_sort, 'count'), plotdir, 'pay_per_person_count',
'Count Per Person')
make_scatter_plot(get_col(p_sort, 'count'),
get_col(p_sort, 'median'),
plotdir,
'pay_per_person_by_count',
'Count',
'Log10 Cost Per Person',
xlim=(-1000, 100000))
plotdir = make_plotdir(plotdir='cms_gender_plots/')
g_group = calc_par_group(df, agg_fns,
['provider_type', 'nppes_provider_gender'],
['pay_per_person'])
g_group = process_by_var(plotdir,
g_group,
col='pay_per_person',
var='nppes_provider_gender')
def calc_by_states(df):
"calc parameters by state"
plotdir = make_plotdir(plotdir='cms_state_service_plots/')
agg_fns = ['count', 'median']
p_group = calc_par_group(df,
agg_fns,
['provider_type', 'nppes_provider_state'],
['pay_per_person', 'pay_per_service'],
print_out=False)
# print('index level provider_types\n', p_group.index.levels[0])
bmap = get_basemap() # read file once for all maps
# minmax = (1.0, 2.7)
# im = p_group.ix['Internal Medicine']['pay_per_service']['median']
# print('%s\n' % provider, im)
# make_state_map(bmap, im, plotdir, 'cost_per_service_internal_medicine', 'Internal Medicine, Median Cost Per Service')
# im = p_group.ix['General Surgery']['pay_per_service']['median']
# print('%s\n' % provider, im)
# make_state_map(bmap, im, plotdir, 'cost_per_service_general_surgery', 'General Surgery, Median Cost Per Service')
# im = p_group.ix['Physical Therapist']['pay_per_service']['median']
# print('%s\n' % provider, im)
# make_state_map(bmap, im, plotdir, 'cost_per_service_physical_therapist', 'Physical Therapist, Median Cost Per Service')
patr = re_compile('[ (/)]+')
# for provider in p_group.index.levels[0]:
# im = p_group.ix[provider]['pay_per_service']['median']
# make_state_map(bmap, im, plotdir, 'cost_per_service_%s' % '_'.join(patr.split(provider.lower())), '%s, Median Cost Per Service' % provider)
plotdir = make_plotdir(plotdir='cms_state_person_plots/')
for provider in p_group.index.levels[0]:
im = p_group.ix[provider]['pay_per_person']['median']
make_state_map(
bmap, im, plotdir,
'cost_per_person_%s' % '_'.join(patr.split(provider.lower())),
'%s, Median Cost Per Person' % provider)
def calc_by_zip(df):
"calc parameters by zip code"
plotdir = make_plotdir(plotdir='cms_zip_person_plots/')
# first validate by state, as a mock of zip
agg_fns = ['count','median']
p_group = calc_par_group(df, agg_fns, ['provider_type','nppes_provider_state'], ['pay_per_person','pay_per_service'], print_out=False)
# print('index level provider_types\n', p_group.index.levels[0])
# bmap = get_basemap() # read file once for all maps
# im = p_group.ix['Cardiology']['pay_per_person']['median']
# make_state_map(bmap, im, plotdir, 'zip_cardiology_per_person', '%s, Median Cost Per Person' % 'Cardiology')
# im = p_group.ix['Cardiology']['pay_per_person']['count']
# make_state_map(bmap, im, plotdir, 'zip_cardiology_per_person_count', '%s, Median Count Per Person' % 'Cardiology')
im_group = p_group.ix['Cardiology']['pay_per_person']
calc_zip_group(im_group, 'Cardiology')
def age_segment_par_groups(df):
"calculate series of grouped age parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_pop_age_plots/')
agg_fns = ['count','sum']
pars = ['beneficiary_age_less_65_count','beneficiary_age_65_74_count','beneficiary_age_75_84_count','beneficiary_age_greater_84_count']
p_group = calc_par_group(df, agg_fns, ['provider_type'], pars)
labels = ['< 65','65-74','75-84','> 84']
age_sums = [ p_group[p]['sum'].sum() for p in pars ]
age_total = 0
for age in age_sums:
age_total += age
print('\nage_range number of patients:')
for label,age in zip(labels, age_sums):
print(' %5s %d (%.2f%%)' % (label, age, 100*age/age_total))
make_group_bar_plots(p_group, 'provider_type', pars, labels, 'sum', 'Patient Age Group', plotdir)
def average_age_par_group(df):
"average age calc per provider type"
df['total_age'] = df['beneficiary_average_age'] * df['total_unique_benes']
# print(df['total_age'])
plotdir = make_plotdir(plotdir='cms_pop_plots/')
agg_fns = ['sum']
p_group = calc_par_group(df, agg_fns, ['provider_type'],
['total_unique_benes', 'total_age'])
p_group['avg_age'] = p_group['total_age']['sum'] / p_group[
'total_unique_benes']['sum']
p_sort = p_group.sort_values(by='avg_age', ascending=False)
print_all_rows(p_sort, ['avg_age'])
make_bar_plot(get_col(p_sort, 'avg_age'),
plotdir,
'beneficiary_average_age',
'Beneficiary Average Age',
xlim=(50, 100))
def calc_by_zip(df):
"calc parameters by zip code"
plotdir = make_plotdir(plotdir='cms_zip_person_plots/')
# first validate by state, as a mock of zip
agg_fns = ['count', 'median']
p_group = calc_par_group(df,
agg_fns,
['provider_type', 'nppes_provider_state'],
['pay_per_person', 'pay_per_service'],
print_out=False)
# print('index level provider_types\n', p_group.index.levels[0])
# bmap = get_basemap() # read file once for all maps
# im = p_group.ix['Cardiology']['pay_per_person']['median']
# make_state_map(bmap, im, plotdir, 'zip_cardiology_per_person', '%s, Median Cost Per Person' % 'Cardiology')
# im = p_group.ix['Cardiology']['pay_per_person']['count']
# make_state_map(bmap, im, plotdir, 'zip_cardiology_per_person_count', '%s, Median Count Per Person' % 'Cardiology')
im_group = p_group.ix['Cardiology']['pay_per_person']
calc_zip_group(im_group, 'Cardiology')
def age_segment_par_groups(df):
"calculate series of grouped age parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_pop_age_plots/')
agg_fns = ['count', 'sum']
pars = [
'beneficiary_age_less_65_count', 'beneficiary_age_65_74_count',
'beneficiary_age_75_84_count', 'beneficiary_age_greater_84_count'
]
p_group = calc_par_group(df, agg_fns, ['provider_type'], pars)
labels = ['< 65', '65-74', '75-84', '> 84']
age_sums = [p_group[p]['sum'].sum() for p in pars]
age_total = 0
for age in age_sums:
age_total += age
print('\nage_range number of patients:')
for label, age in zip(labels, age_sums):
print(' %5s %d (%.2f%%)' % (label, age, 100 * age / age_total))
make_group_bar_plots(p_group, 'provider_type', pars, labels, 'sum',
'Patient Age Group', plotdir)
def pop_calc_par_groups(df):
"calculate series of grouped population parameters, printed by column"
plotdir = make_plotdir(plotdir='cms_pop_plots/')
agg_fns = ['count', 'sum', 'median']
p_group = calc_par_group(df, agg_fns, ['provider_type'], [
'total_unique_benes', 'total_services', 'beneficiary_average_age',
'total_medicare_payment_amt'
])
print('\ntop total_services count'
) # count of number of providers, not patients
p_sort = filter_group_by_var(p_group['total_services'],
agg_fns,
stat='count')
make_bar_plot(get_col(p_sort, 'count', log=True),
plotdir,
'total_services_count',
'Log10 Count Total Services',
xlim=(0.1, 5))
print('\ntop total_services sum')
p_sort = filter_group_by_var(p_group['total_services'],
agg_fns,
stat='sum')
make_bar_plot(get_col(p_sort, 'sum', log=True),
plotdir,
'total_services_sum',
'Log10 Sum Total Services',
xlim=(1, 9))
print('\ntop total_unique_benes sum')
p_sort = filter_group_by_var(p_group['total_unique_benes'],
agg_fns,
stat='sum')
make_bar_plot(get_col(p_sort, 'sum', log=True),
plotdir,
'total_unique_benes_sum',
'Log10 Sum Total Beneficiaries',
xlim=(1, 8))
print('\ntop beneficiary_average_age median')
p_sort = filter_group_by_var(p_group['beneficiary_average_age'],
agg_fns,
stat='median')
make_bar_plot(get_col(p_sort, 'median'),
plotdir,
'beneficiary_average_age_median',
'Median Beneficiary Age Per Provider',
xlim=(50, 100))
print('\ntop total_medicare_payment_amt median')
p_sort = filter_group_by_var(p_group['total_medicare_payment_amt'],
agg_fns,
stat='median')
make_bar_plot(get_col(p_sort, 'median', log=True),
plotdir,
'median_medicare_payment_amt',
'Log10 Median Medicare Payment Amount Per Provider',
xlim=(1, 6))
print('\ntop total_medicare_payment_amt')
p_sort = filter_group_by_var(p_group['total_medicare_payment_amt'],
agg_fns,
stat='sum')
make_bar_plot(get_col(p_sort, 'sum', log=True),
plotdir,
'total_medicare_payment_amt',
'Log10 Total Medicare Payment Amount',
xlim=(1, 10))